Genetic Algorithm: Theory, Literature Review, and Application in Image Reconstruction
Genetic Algorithm (GA) is one of the most well-regarded evolutionary algorithms in the history. This algorithm mimics Darwinian theory of survival of the fittest in nature. This chapter presents the most fundamental concepts, operators, and mathematical m
- PDF / 1,734,023 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 41 Downloads / 238 Views
Abstract Genetic Algorithm (GA) is one of the most well-regarded evolutionary algorithms in the history. This algorithm mimics Darwinian theory of survival of the fittest in nature. This chapter presents the most fundamental concepts, operators, and mathematical models of this algorithm. The most popular improvements in the main component of this algorithm (selection, crossover, and mutation) are given too. The chapter also investigates the application of this technique in the field of image processing. In fact, the GA algorithm is employed to reconstruct a binary image from a completely random image.
1 Introduction One of the fast growing sub-fields of Computational Intelligence is Evolutionary Computation. In this field, there are a large number of algorithms to solve optimization problems. Such algorithms mostly mimic biological evolutionary in nature. The S. Mirjalili (B) · J. Song Dong Institute for Integrated and Intelligent Systems, Griffith University, Nathan, Brisbane, QLD 4111, Australia e-mail: [email protected] J. Song Dong Department of Computer Science, School of Computing, National University of Singapore, Singapore, Singapore e-mail: [email protected] A. S. Sadiq School of Information Technology, Monash University, 47500 Bandar Sunway, Malaysia e-mail: [email protected] H. Faris King Abdullah II School for Information Technology, The University of Jordan, Amman, Jordan e-mail: [email protected] © Springer Nature Switzerland AG 2020 S. Mirjalili et al. (eds.), Nature-Inspired Optimizers, Studies in Computational Intelligence 811, https://doi.org/10.1007/978-3-030-12127-3_5
69
70
S. Mirjalili et al.
framework of most of Evolutionary Algorithms (EAs) are very similar. They start with a population of random solutions. Each solution is evaluated by a fitness function that indicates the suitability of the solutions. Through several iterations, the best solutions are chosen. The best solutions together with stochastic selections are combined to produce the next set of solutions. Evolutionary algorithms are equipped with several random (stochastic) components which select and combine solutions in each population. This makes them unreliable in finding similar solution in each run as opposed to deterministic algorithms. Deterministic algorithms (e.g. brute force search) find the same solution in every run. However, they suffer from slower speed and local solution stagnation when applied to large-scale problems. EAs are stochastic and mostly heuristics. This means to search only a part of search space using heuristics information. Finding the best solutions in each population and using them to improve other solutions allow theses algorithms to search promising regions of the search space instead of all. The wide range of large-scale applications of EAs in recent years show the popularity and flexibility of such techniques. Another advantage of EAs is that they consider optimization problems as block boxes. The shape of search space and its derivation should not be known t
Data Loading...
